ORIGINAL CONTRIBUTION

Transepithelial heme-iron transport: effect of heme oxygenaseoverexpression

M. J. Mendiburo • S. Le Blanc • A. Espinoza •

F. Pizarro • M. Arredondo

Received: 9 March 2010 / Accepted: 2 November 2010 / Published online: 16 November 2010

� Springer-Verlag 2010

Abstract

Background Heme iron is found in the diet mainly in the

form of hemoglobin and myoglobin. It is known that heme

iron (heme-Fe) and inorganic iron are absorbed differently.

Intracellularly, heme oxygenase-1 (HO1) participates in the

cleavage of the heme ring producing biliverdin, CO and

ferrous iron. Iron released from heme becomes part of

labile iron pool, and it can be stored in ferritin or released

through the basolateral membrane. The mechanism by

which heme-Fe is metabolized within cells is not com-

pletely understood.

Objective This study focused on the uptake and transport

of heme iron and on the role of heme oxygenase-1 on heme

iron metabolism.

Design Caco-2 cells were incubated with different con-

centrations of heme-Fe. A full-length heme oxygenase-1

cDNA was expressed in Caco-2 cells and intracellular iron

and heme-Fe content, heme uptake, heme and iron trans-

port and heme oxygenase-1 immunolocalization were

assessed in these cells.

Results Heme-Fe was bioavailable and induced an intra-

cellular increase in iron, ferritin and HO1 levels and a

decrease in DMT1 expression. In cells overexpressing

HO1, heme-Fe uptake and transepithelial Fe transport was

higher than in controls. Most heme-Fe was metabolized to

free iron, most of which was found mainly in the basolat-

eral chamber. However, there is a fraction of heme that is

delivered intact to the basolateral side. In a high heme-Fe

condition, HO1 is found near the plasma membrane.

Conclusions These results suggest that heme oxygenase-

1 catabolizes most of the heme-Fe and favors iron influx

and efflux in intestinal cells.

Keywords Heme-Fe � Transepithelial transport �Caco-2 cells � Heme oxygenase

Introduction

One of the defense mechanisms most widely used in nature

is enzyme heme oxygenase-1 (HO-1). This microsomal

enzyme performs the seemingly lackluster function of

catabolizing heme to generate bilirubin (an antioxidant),

carbon monoxide, and free iron (a potent pro-oxidant) [37].

Three heme oxygenase (HO) isoforms have been identified,

HO-1, HO-2, and HO-3. HO-1 is a 32-kDa heat shock

protein, which is inducible by numerous noxious stimuli.

The common characteristic of many of these inducers is

their ability to cause oxidative stress. These include, but are

not limited to: heme and heavy metals [14], hyperoxia [24],

hypoxia [7, 25], UV light, hydrogen peroxide [19, 23],

lipopolysaccharide [5], hyperthermia [13] and endotoxin

[6]. HO-1 expression is primarily regulated at the tran-

scriptional level [1, 9, 11, 15]. HO-2 is a constitutively

synthesized 36-kDa protein, which is abundant in brain and

testis [36, 39]. The third isoform, HO-3, has been reported

as a pseudogene derived from HO-2 transcripts [17].

Remarkably, the activity of this enzyme results in pro-

found changes in the ability of cells to protect themselves

against oxidative injury. HO-1 has been shown to have

anti-inflammatory, anti-apoptotic, and anti-proliferative

effects, and it is now known to have salutary effects in

diseases as diverse as atherosclerosis and sepsis. Stocker

[35] has proposed that this enzyme might provide cellular

M. J. Mendiburo � S. Le Blanc � A. Espinoza � F. Pizarro �M. Arredondo (&)

Micronutrient Laboratory, Nutrition Institute and Food

Technology, El Lıbano 5524, Macul, Santiago, Chile

e-mail: marredon@inta.cl

123

Eur J Nutr (2011) 50:363–371

DOI 10.1007/s00394-010-0144-5

protection. The mechanism by which this enzyme confers

cellular protection is only beginning to be unraveled. The

appeal is readily apparent: if we can understand how cells

are able to protect themselves from oxidative stress, then

our understanding and ability to intervene in disease pro-

cesses will be immeasurably advanced [28].

Both HO and its substrate, heme, are highly conserved

molecules across almost all forms of life, from algae to

mammals. Molecules so evolutionarily conserved and

ubiquitous generally serve a necessary and fundamental

purpose [28]. There are relatively few studies describing

the mechanism of intestinal heme iron absorption despite

the importance of heme iron as a highly bioavailable source

of dietary iron. Populations that consume meat as a sig-

nificant component of their diet are normally iron replete.

In fact, it has been determined that two-thirds of absorbed

dietary iron in North America and Europe is derived from

heme iron, although it only comprises one-third of dietary

iron. Intestinal absorption of heme iron is higher than that

of non-heme iron, suggesting that heme may be a preferred

iron source in iron deficiency; it may also be a source of

dietary iron to avoid when iron status is high, such as in

hemochromatosis [31]. In this study, we determined the

effect of HO1 over-expression on heme-Fe bioavailability

and intracellular iron transport.

Materials and methods

Cell culture

Caco-2 cells (American Type Culture Collection HTB37

Rockville, MD) (1 9 105 cells) were cultured at 37 �C and

5% CO2 in 25-cm2 flasks with Iscove0s media (Gibco Life

technologies, Grand Island, NY) supplemented with 10%

FBS, 10 kU/L penicillin/streptomycin, and 25 mg/L fun-

gizone (Gibco). The cells were trypsinized and re-seeded in

either 24- or in 6-well plates (0.2–0.5 9 105 cells) or on

polycarbonate membranes of 0.33 lM pore size and

6.5 mm diameter (0.2 9 105 cells) (Transwells, Corning,

Costar, Cambridge, MA). The medium was changed every

2–3 days.

Isotopic labeling and digestion of hemoglobin (Hb)

An iron isotope (55Fe) of high specific activity was used as

a tracer (NEN, Life Science Products, Boston, MA).

Labeled hemoglobin (Hb) was prepared from red blood

cells obtained from New Zealand rabbits that had received

an intravenous injection of 74 MBq of 55Fe as ferric citrate

diluted in saline solution. The rabbits were bled through a

cardiac puncture 15 days later. The radioactive red blood

cells were centrifuged (1,0009g for 15 min at 22 �C) and

washed with saline solution, then hemolyzed by freezing

and dehydrated by lyophilization. The cell extract was

labeled with a specific activity of 2,460 kBq of 55Fe per mg

of heme-Fe obtained. Partial digestion of Hb solution was

performed. Briefly, Hb solution containing 2 mM 55Fe or56Fe were digested with 0.1% pepsin at pH 2.0 for 1 h at

37 �C. This solution was diluted by adding HEPES buffer

(pH 7.2) to increase the pH to 6.8. A digestion of 52 ± 3%

was estimated in the Hb-digest by measuring Hb content in

the supernatant, which was used as a source of heme-Fe.

Elemental Fe was determined by atomic absorption

spectrometry with graphite furnace (SIMAA 6100, Perkin-

Elmer, Norwalk, CT).

Caco-2 cells with different heme or non-heme iron

concentrations

Caco-2 cells were grown for 7 days in the presence of

heme-Fe, as described earlier. The incubations were made

using a stock heme-Fe solution with the following con-

centrations: 0.1, 5, 10, 20, and 50 lM of heme-Fe. The

medium was changed every 2–3 days. After 7 days, the

cells were trypsinized and re-seeded (5–10% initial

cells = 1 9 105), as previously described, for another

7 days. After the treatment, a cell lysate was prepared with

lysis buffer (in mM: 10 HEPES, pH 7.5, 3 MgCl2, 40 KCl,

1 PMSF, 1 DTT, and 5% glycerol, 0.5% Triton X-100 and

1 9 protease inhibitor cocktail (Sigma, St Louis, MO). The

mix was incubated for 15 min on ice and centrifuged for

10 min at 4 �C and 15,000 rpm. The supernatant was ali-

quoted and stored at -70 �C. Protein concentration was

determined by Lowry method [26], intracellular ferritin by

ELISA (rabbit anti-human ferritin code A0133 and per-

oxidase-conjugated rabbit anti-human ferritin code P0145,

Dako Corp, Denmark) and total iron by spectrometric

atomic absorption with graphite furnace (SIMAA 6100,

Perkin-Elmer, Norwalk, CT).

Antibodies and immunodetections

Western blotting assays were performed on cell lysate to

study the expression of HO-1, DMT1 (Divalent Metal

Transporter 1) and Ireg1 (also, ferroportin). Fifty micro-

grams of cell lysate were loaded and separated on 14%

(HO-1) and 8% (DMT1 and Ireg1) SDS–PAGE, and trans-

ferred to a nitrocellulose membrane. The primary antibody

for HO1 was a rabbit polyclonal antibody (Santa Cruz Bio-

technology, Inc). DMT1 and Ireg1 antibodies were provided

by Dr. MT Nunez, Faculty of Science, University of Chile.

DMT1 and Ireg1 antibodies were rabbit polyclonal anti-

bodies prepared against COOH-terminal peptide. The sec-

ondary antibody was peroxidase-labeled goat anti-rabbit

immunoglobulin G (Sigma Chemical). For membrane

364 Eur J Nutr (2011) 50:363–371

123

examination, the enhanced chemiluminescence Western

blotting detection system (Amersham, Arlington Heights,

IL) was used. Membranes were stripped with 100 mM citric

acid (pH 3.0) and then re-blotted with anti-actin (Sigma

Chemical).

Subcellular localization of HO1 in Caco-2 cells

Cells were grown in polyester membrane Transwells

(Costar) for 14 days, incubated with or without 50 lM

heme-Fe for the last 5 days, then fixed with 4% parafor-

maldehyde, permeated with 0.2% Triton X-100 in saline,

and reacted with anti-HO1 antibody. A second antibody

was FITC-labeled anti-rabbit IgG antibody (Sigma Chem.

Co.). Fluorescence was determined in a Zeiss MP40 con-

focal microscope. For co-immunolocalization analysis,

cells were incubated over night with mouse anti-HO1

(1:250, US Biological) and rabbit anti-Glut1 (1:200, US

Biological). Then, cells were washed 69 with PBS-BSA

for 5 min and then incubated with Alexa 546 anti-mouse

IgG (1:500, Molecular Probes) and Alexa 488 anti-rabbit

IgG (1:500, Molecular Probes). Fluorescence was deter-

mined as mentioned earlier.

Vector construction and transfection of Caco-2 cells

with ho1 cDNA

Total RNA was isolated from Caco-2 cells with Trizol

reagent (Invitrogen) according to manufacturer instruc-

tions. Briefly, Caco-2 cells seeded in 25-cm2 bottles were

cultured for 7 days and lysed with 2.5 mL of Trizol. RNA

was resuspended in DEPC water, aliquoted and stored at

-80 �C. cDNA was obtained by reverse transcription. The

reaction contained 5 lg of RNA and 0.5 lg oligo-dT. The

mix was incubated at 70 �C for 10 min and then at 4 �C for

1 min. Two microliters of 109 PCR buffer, 1 lL of

50 mM MgCl2, 1 lL of 10 mM dNTPs, and 2 lL of 0,1 M

DTT were added to a final volume of 20 lL and incubated

at 42 �C for 5 min. Then 200 U of MMLV Reverse

Transcriptase (Invitrogen Corporation, Carlsbad, CA,

USA) was added, and the mix was incubated at 42 �C for

50 min. The reaction was stopped by incubation at 70 �C

for 15 min. One microliter of RNase H (Invitrogen) was

added, and the mix was incubated at 37 �C for 20 min.

ho1 full-length cDNA (GenBank accession number:

NM_002133) was amplified by PCR using the following

primers: HO1s 50-GAACGAGCCAAGCTTCGGCCGGAT

G-30 (position 59-83) and HO1a 50-GGAGCCAGCGC

GGCCGCATACACAT-30 (position 942-966). Underlined

letters indicate nucleotide changes with respect to the ho1

mRNA sequence in order to introduce restriction sites for

HindIII and NotI. The cDNA was used as a template

for PCR amplification using the following cycles: 94 �C

for 35 min, 94 �C for 15 seg, 60 �C for 1 min, 72 �C for

1 min, and 72 �C for 10 min. A band of 908 pb was

obtained. The PCR product was digested with HindIII and

NotI and purified with Gene Clean II Bio 101 using silica

(Sigma–Aldrich, St Louis, MO). ho1 cDNA was cloned in

the pcDNA3.1myc-his (Invitrogen) expression vector.

E. coli DH5a were transformed, and positive clones were

verified with restriction analysis and sequencing. Caco-2

cells were grown at subconfluence (50–70%) and trans-

fected with pcDNA3.1myc-his-ho1 vector (HO1 cells) or

pcDNA3.1myc-his (control cells) using Lipofectamin

2000 (Invitrogen). Transfected cells were selected using

800 lg/mL G418 (Gibco) for 48 h, and then the cells were

maintained in Iscove0s media as earlier with 400 lg/mL

G418.

Intracellular total levels of iron and ferritin

HO-1 and control cells were grown in 12-well plates in

selection medium for 7 days. A cell lysate was prepared,

then digested with 65% nitric acid (1:1) and incubated at

60 �C overnight. Total iron was determined by spectro-

metric atomic absorption with graphite furnace Simaa 6100

(Perkin Elmer). Intracellular ferritin was determined in cell

lysate using ELISA (rabbit anti-human ferritin Code A0133

and peroxidase-conjugated rabbit anti-human ferritin code

P0145, Dako Corp, Denmark).

Heme-Fe and iron uptake and transport

HO1 and control cells were plated onto 0.33 cm2 poly-

carbonate inserts for 12 days and grown as previously

described. The medium was changed every 3 days. Inserts

were used when they attained stable resistance values

between 250 and 280 X cm2. On the day of the experiment,

the cells were washed with 19 PBS, and 50 lM heme-55Fe

or 25 lM 55FeCl3 (Fe:ascorbic acid 1:5) was added to the

apical side in transport buffer (in mM: 50 MOPS-Na; 94

NaCl; 7.4 KCl; 0.74 MgCl2; 1.5 CaCl2; 5 glucose, pH 6.5)

at different times (0–60 min) at 37 �C. The reaction was

stopped washing the inserts 3 times with cold PBS/1 mM

EDTA. Afterward the membranes were cut out and 1 mL

of scintillation liquid was added to each tube. Also, 100 lL

of the basolateral medium was diluted with scintillation

liquid. The radioactivity from 55Fe in both the membrane

and the basolateral medium was measured in a gamma

counter (Beckman LS 5000 TD).

Determination of non-heme iron and protoporphyrin

transport at the basolateral side

For protoporphyrin transport determination, HO1 and con-

trol cells were grown in bicameral chambers (0.66 cm2) in

Eur J Nutr (2011) 50:363–371 365

123

Iscove0s media with 10% low-Fe FBS and G418

(400 lg/mL). On the day of the experiment, 50 lM heme-Fe

was added to the apical chamber, and the cells were incu-

bated at 37 �C for different periods of time (0–60 min). The

transepithelial electric resistance (TEER) was monitored

during the experiment. Inserts with TEER lower than

240 X cm2 were eliminated. Basolateral media was col-

lected and protoporphyrin concentrations were measured in

a Shimadzu UV-1601 spectrophotometer at 398 nm. We

used a molar extinction coefficient of e = 1.56 9 105 M-1

cm-1 for this calculation. For non-heme iron transport

determination, control and HO1 cells were grown in

bicameral inserts (1 cm2). On the day of the experiment,

inserts were washed with MOPS-saline buffer. Afterward,

0.1 lM calcein in MOPS-glucose buffer and MOPS-glucose

were added to the basolateral and apical chamber, respec-

tively. Calcein fluorescence was measured, then heme-Fe

(10 lM) was added to the apical compartment and the

decrease in fluorescence was measured again in 30 cycles of

1 min each. Finally, 10 lL of 10 lM SIH were added to

chelate Fe.

Heme oxygenase enzymatic activity

A cell lysate from control and HO1 cells was prepared using

a non-denaturing lysis buffer (20 mM Tris–HCl; pH 7.4;

0.5% Triton X-100; protein inhibitor cocktail). Two hundred

fifty micrograms of cell lysate were incubated with 600 lL

of B buffer (100 mM KH2PO4, pH 7.4), 100 lL of 150 lM

hemin, 100 lL of 100 lg/mL rat liver extract containing

biliverdin reductase and 100 lL of 10 mM NADPH for 1 h

at 37 �C in the dark. Bilirubin formed in the reaction was

extracted with 1 mL of chloroform for 1 h at room temper-

ature in a shaker (100 rpm). Then, absorbance was measured

at 530 nm (molar extinction coefficient of bilirubin:

43.5 mM-1 cm-1). HO enzymatic activity was expressed as

nmole of bilirubin/mg protein/hr.

Bilirubin reductase isolation

Rat livers (Rattus norvegicus) were perfunded in situ with

saline (0.9% NaCl, pH 7.2) until complete discoloration,

dissected, homogenized in lysis buffer A (0.1 M sodium

citrate, pH 5.0; 10% glycerol) and centrifuged for 20 min

at 10,0009g and 1 h at 105,0009g. The supernatant was

diluted in 20 mM KH2PO5; 135 mM KCl; 0.1 mM EDTA;

pH 7.4. Protein concentration was determined. The extract

was aliquoted and stored at -20 �C.

Statistical analysis

Variables were tested in triplicate, and the experiments were

repeated at least twice. Variability among experiments was

\20%. One-way ANOVA and T test were used to test dif-

ferences in mean values, and Bonferroni0s post hoc test was

used for comparisons (SAS 8.0E, SAS Institute Inc., Cory,

NC). Differences were considered significant if p \ 0.05.

Results

Intracellular iron and ferritin content in Caco-2 cells

incubated with heme-Fe

To determine the bioavailability of heme-Fe, we measured

total intracellular iron (Fig. 1a) and ferritin (Fig. 1b) in

Caco-2 cells incubated with different extracellular heme-Fe

concentrations for two passages. Intracellular Fe increased

(2.8 ± 0.4–10.8 ± 0.8 nmole Fe/mg protein) when extra-

cellular heme-Fe increased from 0.5 to 100 lM (one-way

ANOVA: p \ 0.001). Intracellular ferritin also increased

(1.3 ± 0.2–43.6 ± 0.6 nmole Fn/mg protein) in the same

range of extracellular heme-Fe (one-way ANOVA:

p \ 0.001). HO1 expression was induced at high extra-

cellular heme-Fe concentrations (p \ 0.01). As expected,

DMT1 expression decreased when intracellular Fe

increased (p \ 0.05). No change was observed in Ireg1

expression (Fig. 1c, d).

Immunolocalization of heme oxygenase

To determine the intracellular localization of HO1 enzyme,

Caco-2 cells were incubated with or without 50 lM heme-

Fe for 5 days and subjected to confocal microscopy. In

Caco-2 cells incubated with heme-Fe, HO1 changed its

localization from perinuclear to a domain close to the

plasma membrane (Fig. 2A:a) compared with control cells

(Fig. 2A:c). To confirm this result, a co-immunolocaliza-

tion was performed in cells incubated with GLUT1 (2B:a)

and HO1 (2B:b) antibodies. We observed that HO1 co-

localizated with GLUT1 transporter (a basolateral marker

in Caco-2 cells), which suggests that HO1 could be asso-

ciated to an inner plasma membrane domain.

Characterization of Caco-2 cells over-expressing HO1

To enhance the expression of HO1, Caco-2 cells were

transfected with pcDNA3.1myc-his-ho1 vector (HO cells),

and HO1 overexpression was confirmed by Western blot-

ting (Fig. 3a). Under this condition, HO1 enzymatic

activity increased from 6.4 ± 2.1 to 10.2 ± 0.4 nmole

bilirubin/hr/mg protein, in control and HO1 cells, respec-

tively (p \ 0.05; Fig. 3b). Intracellular ferritin concentra-

tion also increased from 0.9 ± 0.1 to 6.4 ± 1.9 ng ferritin/

mg protein, in control and HO1 cells, respectively

(p \ 0.01, Fig. 3c).

366 Eur J Nutr (2011) 50:363–371

123

Heme-Fe and non-heme iron uptake and transport

in HO cells

To determine the effect of HO1 over-expression on heme-

Fe and iron uptake, HO1 cells seeded in bicameral inserts

were incubated with 50 lM heme-55Fe or 25 lM 55FeCl3added to the apical side, and radioactivity from the cell

lysate and basolateral medium was measured to determine

heme-Fe or iron uptake and iron transport, respectively.

HO1 cells showed an increase in heme-Fe uptake

(p \ 0.02) (Fig. 4a) and transport of 55Fe to the basolateral

side, compared with control cells (p \ 0.01) (Fig. 4b).

There were no significant changes in apical non-heme iron

uptake or apical to basolateral iron transport.

Apical to basolateral protoporphyrin and iron transport

in HO cells

To elucidate which form of iron (i.e. heme-Fe or ferrous

Fe) is the main contributor of apical to basolateral 55Fe

transport, heme (as a protoporphyrin) and non-heme iron

were measured from the basolateral side after HO1 cells

were incubated with 50 lM heme-Fe apically. Transport of

protoporphyrin was significantly lower in HO1 cells com-

pared to control cells (p \ 0.001) (Fig. 5a). However, iron

transport to the basolateral side, measured by calcein

quenching, was higher in HO1 cells than in control cells

(one-way ANOVA: p \ 0.02) (Fig. 5b).

Discussion

The process of non-heme-Fe absorption by enterocytes is

very well known [4, 10, 30, 34]. However, there are few

studies regarding heme-Fe uptake and transport by intes-

tinal cells. The movement of heme into and within cells

was thought to occur by diffusion. However, the chemical

properties of heme make diffusion too slow to keep pace

with biological processes [21]. It has been suggested that

heme enters cells as an intact molecule of metalloporphyrin

[33], and three different mechanisms have been proposed

for heme uptake: (1) pinocytosis [29, 42], (2) heme

receptor on duodenal brush border [16] and (3) via the

heme transporter HCP1 (heme carrier protein 1) [32],

whose activity should be closely related to heme oxygenase

enzyme. Iron released from heme is later found in the

blood [10, 40, 41]. However, the mechanism of intracel-

lular heme movement from apical to basolateral side has

yet to be explained.

Heme-Fe was bioavailable for Caco-2 cells when they

were incubated with different heme-Fe concentrations.

Similar to what is observed in Caco-2 cells incubated with

Intracellular Fe

-1

10

12

Intracellular Ferritin

-1

30

40**

nmol

e F

e* m

g pr

otei

n2

4

6

nmol

e F

erri

tin

*mg

prot

ein

10

20

Extracelular Heme-Fe, uM0 10 50 100

Extracelular Heme-Fe, uM0 10 50 100

*

HO1

Actin

0 10 20 40 60 100

DMT1

Heme-Fe,µM Heme-Fe,µM

1.4

1.6

1.810204060

100

*

Actin

DMT1

Ireg1

Actin

prot

ein/

acti

n, A

U

0.8

1.2

1.0

*HO1 DMT1 IREG1

8

A B

C D

Fig. 1 Caco-2 cells were

incubated with different heme-

Fe concentration (range

0–100 lM). Intracellular total

Fe was measured by

spectrometric atomic absorption

with graphite furnace Simaa

6100 (a), ferritin by ELISA (b),

Western blot of HO1, DMT1,

and Ireg1 (c) and densitometric

analysis of C (d). (*One-way

ANOVA, p \ 0.001)

Eur J Nutr (2011) 50:363–371 367

123

non-heme iron [2], in cells incubated with heme-Fe, total

intracellular iron and intracellular ferritin concentration

increased. Furthermore, HO1 protein expression was

dependent of heme-Fe bioavailability, as we had previously

shown [3]. As a result of the increased intracellular Fe,

DMT1 transporter expression decreased. However, Ireg1

protein expression did not change in the present experi-

mental conditions. These results indicate that heme-Fe was

available for the cell and that once iron is released it

becomes part of the intracellular Fe pool. Similar results

have been shown by Eisenstein et al. [12], who observed

that release of iron from heme is necessary for maximal

induction of ferritin synthesis and that direct donation of

iron to the intracellular iron pool induced ferritin synthesis

significantly, but it was not a good inducer of HO. In

humans, Pizarro et al. [31], demonstrated that heme-Fe

absorption is a saturable process post-ingestion of physio-

logical doses of either hemoglobin or myoglobin.

To determine the role of HO1 enzyme on intracellular

iron transport, we transfected Caco-2 cells with the HO1

cDNA. We observed an increase in HO1 protein expression

and enzymatic activity and in intracellular ferritin con-

centration in these cells. The increase in ferritin concen-

tration is an indicator of increased iron availability. Under

Fig. 2 Cellular HO1

distribution in Caco-2 cells.

a Caco-2 cells were reacted with

anti-HO1 antibody followed by

FITC-labeled anti-rabbit IgG

antibody as described in

Methods (A and C). Caco-2 cells

were incubated with (A) or

without (B) 50 lM heme-Fe for

5 days. HO1 localization was

assessed following FITC

fluorescence in a confocal

microscope. Phase contrast of

Caco-2 cells (B and D)

(b) Co-immunolocalization of

HO1 and GLUT1 transporter.

Caco-2 cells were incubated

with rabbit anti-GLUT1 (A) and

mouse anti-HO1 (B). Then cells

were incubated with Alexa 546

anti-mouse IgG and Alexa 488

anti-rabbit IgG. Fluorescence

was determined as mentioned

earlier

368 Eur J Nutr (2011) 50:363–371

123

these conditions, heme-Fe uptake by HO1 cells was

increased, which correlated positively with iron availabil-

ity. This result suggests that in HO1 cells, which show a

high HO1 enzymatic activity, the catabolism of intracel-

lular heme is enhanced, resulting in a decrease in the

intracellular heme/non-heme-Fe ratio. This is an indication

that HO1 cells have a higher activity than controls cells

that results in a decrease in heme transport and an increase

in iron transport. Furthermore, the over-expression of HO1

triggers an increase in heme uptake, but does not modify

non-heme-Fe uptake, which also suggests that the intra-

cellular heme levels in these cells are lower, leading to a

compensatory increment in heme uptake. The latter is

possibly due to an up-regulation of the expression of a

heme importer. Taken together, these results suggest that

the increase in intracellular ferritin is due to an increase in

heme uptake. In HO cells, heme-Fe uptake and apical to

basolateral iron transport were higher than in control cells.

Independent of iron concentration, the cells exposed to

heme-Fe transported iron out of the cell at a higher rate.

Enzymatic activity of HO

*

-1

88

Intracellular Ferritin

-1

0.06

0.08

0.10

-1

4

6

4

6

**HO1

Actin

HO1Control

nmol

e bi

lirru

bin

* m

g pr

otei

n

0.02

0.042

Control

2

* h

Control HO1

ng F

erri

tin

* m

g pr

otei

n

HO1

A B C

Fig. 3 Caco-2 cells were transfected with HO1 cDNA (HO1 cells). a HO1 Western blot in HO1 cells and control cells; b heme oxygenase

activity; and c intracellular ferritin concentration. (T test: *p \ 0.01; **p \ 0.001)

300ControlHO1

Apical Heme-

100ControlHO1

Apical to Basolateral 55 Fe transport

*

150

200

250 *

60

80

-1

50

100

20

40

mg

prot

ein

600

800

80

100

mol

e 55F

e *

Control

HO1Control

HO1

40040

60

p

-1 m

g pr

otei

nm

ole 55

Fe *

p

10 20 30 40 50 60Time, min

10 20 30 40 50 60

20

Time, min Apical

55Fe uptake Apical to Basolateral

55Fe transport

55Fe uptake

200

A

C

B

D

Fig. 4 Heme-Fe uptake (a) and

apical to basolateral Fe transport

(b) in HO1 cells and control

cells. Caco-2 cells were

incubated at 37 �C for 0–60 min

with 10 lM heme-55Fe.

Radioactivity was measured in

membranes and in basolateral

media, (*two-way ANOVA,

p \ 0.001)

Eur J Nutr (2011) 50:363–371 369

123

It has been proposed that cells exposed to heme-Fe cannot

sense iron uptake [8, 38]. However, as we were following55Fe, we cannot discriminate which form of iron (heme-Fe

or Fe) was transported to the basolateral side.

We also studied whether heme is transported to the

basolateral side. The results suggest that most of the heme-

Fe was catabolized in the HO1 cells. However, a propor-

tion of heme-Fe (as protoporphyrin) is transported intact to

the basolateral side (control cells). Free heme and proto-

porphyrin are toxic to the cell; therefore, cells must balance

their intracellular metabolism, and for this reason free

heme must be transported out of the cells. This transport

could be performed by FLVCR (Feline Leukemia Virus

subgroup C Receptor) [18]. FLVCR protects erythroid cells

from heme toxicity during differentiation. This heme-efflux

protein is expressed in other cells and tissues, including the

intestine, where they appear to function as apical/basolat-

eral heme exporters to prevent toxicity within the entero-

cytes [22, 32]. In this study, iron in the basolateral media

was threefold higher in HO1 cells than in control cells.

Heme oxygenase-1 distribution in control cells was

mainly perinuclear, which corroborates previous results

from this group [27]. However, in cells incubated with heme-

Fe, the expression of HO1 was higher and detected at a

peripheral compartment. In Caco-2 cells that were over-

expressing HO1, the intracellular localization of HO1

changed from a perinuclear to a putative plasma membrane

topology. Kim et al. [20], using inducers of HO1 or over-

expression of HO1, demonstrated an increase in HO1 protein

in a detergent-resistant fraction containing caveolin-1.

Inducible HO activity appeared in plasma membrane,

cytosol, and isolated caveolae. HO1-GLUT1 co-localization

suggests a basolateral localization of HO1 in Caco-2 cells.

However, it is necessary to take into account that most of the

membrane surface in these cells corresponds to a basolateral

membrane. It is probably that HO1 is associated to a struc-

ture that itself interacts with the plasma membrane. Further

experiments are necessary to dilucidate this question.

In summary, our study shows that Caco-2 cells can be

used as a model to study intestinal heme-iron metabolism

when high specific activity heme 55Fe is used. Heme-Fe is

taken up by the cells, mostly degraded by HO1, and free

iron forms part of the labile iron pool, which is delivered

either to ferritin or to the basolateral side. A fraction of

heme can be transported out of the cells intact.

Acknowledgments This work was supported by Fondo Nacional de

Ciencia y Tecnologıa, grant 1085173 to MAO.

References

1. Agarwal A, Shiraishi F, Visner GA, Nick HS (1998) Linoleyl

hydroperoxide transcriptionally upregulates heme oxygenase-1

gene expression in human renal epithelial and aortic endothelial

cells. J Am Soc Nephrol 9:1990–1997

2. Arredondo M, Orellana A, Nunez MT (1997) Intracellular iron

regulates iron regulatory protein activity and iron absorption in

intestinal epithelial (Caco-2) cells. Am J Physiol 273(36):G275–

G280

3. Arredondo M, KIoosterman J, Nunez S, Segovia F, Candia V,

Flores S, Le Blanc S, Olivares M, Pizarro P (2008) Heme iron

uptake by Caco-2 cells is a saturable, temperature sensitive and

modulated by extracellular pH and potassium. Biol Trace Elem

Res 125:109–119

4. Beard JL, Dawson H, Pinero DJ (1996) Iron metabolism: a

comprehensive review. Nutr Rev 54:295–317

Fe FeFe

Fe Fe

Fe

Fe

Fe

398 nm

-1 Control

Fe

CalceinSIH

Fe

Fe -SIHFeFe

300

400

500 HO1

80

100SIH

Heme-Fe

*

100

200

Flu

ores

ccen

ce, %

40

60

Control *

Apical-basolateral heme transportTime, min

pmol

e pr

otop

oorp

hyri

n *

ml

10 20 30 40 50 60

Time, min0 10 20 30 40

HO1

Apical-basolateral Fe transport

Fe Fe

A BFig. 5 Apical to basolateral

heme transport (a) and apical to

basolateral Fe transport. Caco-2

cells were cultured in bicameral

chambers and incubated with

heme-Fe in the apical side.

Heme (a) and Fe (b) transport

was measured in the basolateral

chamber. (*Two-way ANOVA,

p \ 0.001)

370 Eur J Nutr (2011) 50:363–371

123

5. Camhi SL, Alam J, Otterbein L, Sylvester SL, Choi AM (1995)

Induction of heme oxygenase-1 gene expression by lipopolysac-

charide is mediated by AP-1 activation. Am J Respir Cell Mol

Biol 130:387–398

6. Carraway MS, Ghio AJ, Taylor JL, Piantadosi CA (1998)

Induction of ferritin and heme oxygenase-1 by endotoxin in the

lung. Am J Physiol Lung Cell Mol Physiol 275:L583–L592

7. Carraway MS, Ghio AJ, Carter JD, Piantadosi CA (2000)

Expression of heme oxygenase-1 in the lung in chronic hypoxia.

Am J Respir Cell Mol Biol 278:L806–L812

8. Cermak J, Baila J, Jacob HS, Baila G, Enright H, Nath K,

Vercellotti GM (1993) Tumor cell heme uptake induces ferritin

synthesis resulting in altered oxidant sensibility: possible role in

chemotherapy efficacy. Cancer Res 53:5308–5313

9. Choi AMK, Alam J (1996) Heme oxygenase-1: function, regu-

lation, and implication of a novel stress-inducible protein in

oxidant-induced lung injury. Am J Respir Cell Mol Biol 15:9–19

10. Conrad ME, Umbreit JN (2000) Iron absorption and transport—

an update. Am J Hematol 64:287–298

11. Durante W, Kroll MH, Christodoulides N, Peyton KJ, Schafer AI

(1997) Nitric oxide induces heme oxygenase-1 gene expression

and carbon monoxide production in vascular smooth muscle

cells. Circ Res 80:557–564

12. Eisenstein RS, Garcia-Mayolt D, Pettingellt W, Munro HM (1991)

Regulation of ferritin and heme oxygenase synthesis in rat

fibroblasts by different forms of iron translational control/iron-

regulated proteins/metabolic regulation/iron regulatory element-

binding protein/chelatable iron). Proc Nat Acad Sci USA

88:688–692

13. Ewing JF, Maines MD (1991) Rapid induction of heme oxy-

genase-1 mRNA and protein by hyperthermia in rat brain: heme

oxygenase 2 is not a heat shock protein. Proc Natl Acad Sci USA

88:5364–5368

14. Eyssen-Hernandez R, Ladoux A, Frelin C (1996) Differential

regulation of cardiac heme oxygenase-1 and vascular endothelial

growth factor mRNA expressions by hemin, heavy metals, heat

shock and anoxia. FEBS Lett 382:229–233

15. Fogg S, Agarwal A, Nick HS, Visner GA (1999) Iron regulates

hyperoxia-dependent human heme oxygenase-1 gene expression

in pulmonary endothelial cells. Am J Respir Cell Mol Biol

20:797–804

16. Grasbeck R, Majuri R, Kouvonen I, Tenhunen R (1982) Spectral

and other studies on the intestinal haem receptor of the pig.

Biochem Biophys Acta 700:137–142

17. Hayashi S, Omata Y, Sakamoto H, Higashimoto Y, Hara T,

Sagara Y, Noguchi M (2004) Characterization of rat heme oxy-

genase-3 gene Implication of processed pseudogenes derived

from heme oxygenase-2 gene. Gene 336(2):241–250

18. Keel SB, Doty RT, Yang Z, Quigley JG, Chen J, Knoblaugh S,

Kingsley PD, De Domenico I, Vaughn MB, Kaplan J, Palis J,

Abkowitz JL (2008) A heme export protein is required for red blood

cell differentiation and iron homeostasis. Science 319:825–828

19. Keyse SM, Tyrrell RM (1987) Both near ultraviolet radiation and

the oxidizing agent hydrogen peroxide induce a 32-kDa stress

protein in normal human skin fibroblasts. J Biol Chem

262:14821–14825

20. Kim HP, Wang X, Galbiati F, Ryter SW, Choi AM (2004)

Caveolae compartmentalization of heme oxygenase-1 in endo-

thelial cells. FASEB J 18:1080–1089

21. Krishnamurthy P, Xie T, Schuetz JD (2007) The role of trans-

porters in cellular heme and porphyrin homeostasis. Pharmacol

Ther 114:345–358

22. Latunde-Dada GO, Simpson RJ, McKie AT (2006) Recent

advances in mammalian haem transport. Trends Biochem Sci

31:182–188

23. Lautier D, Luscher P, Tyrrell RM (1992) Endogenous glutathione

levels modulate both constitutive and UVA radiation/hydrogen

peroxide inducible expression of the human heme oxygenase

gene. Carcinogenesis 13:227–232

24. Lee PJ, Alam J, Wiegand GW, Choi AM (1996) Overexpression

of heme oxygenase-1 expression in human pulmonary epithelial

cells results in cell growth arrest and increased resistance to

hyperoxia. Proc Natl Acad Sci USA 93:10393–10398

25. Lee PJ, Jiang BH, Chin BY, Iyer NV, Alam J, Semenza GL, Choi

AM (1997) Hypoxia-inducible factor-1 mediates transcriptional

activation of the heme oxygenase-1 gene in response to hypoxia.

J Biol Chem 272:5375–5381

26. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ (1951) Protein

measurement with the Folin phenol reagent. J Biol Chem

193:265–275

27. Mendiburo MJ, Flores S, Pizarro F, Arredondo M (2006) Heme

oxygenase-1 overexpression increases iron fluxes in Caco-2 cells.

Biol Res 39:195–197

28. Morse D, Choi MAK (2002) Heme oxygenase-1 the ‘‘emerging

molecule’’ has arrived. Am J Respir Cell Mol Biol 27:8–16

29. Parmley RT, Barton JC, Conrad ME, Austin RL, Holland RM

(1981) Ultraestructural cytochemistry and radioautography of

hemoglobin-iron absorption. Exp Mol Pathol 34:131–134

30. Philpott CC (2002) Molecular aspects of iron absorption:

insights into the role of HFE in hemochromatosis. Hepatology

35:993–1001

31. Pizarro F, Olivares M, Hertrampf E, Mazariegos DI, Arredondo

M (2003) Heme-iron absorption is saturable by heme-iron dose in

women. J Nutr 133:2214–2217

32. Quigley JG, Yang Z, Worthington MT, Phillips JD, Sabo KM,

Sabath DE, Berg CL, Sassa S, Wood BL, Abkowitz JL (2004)

Identification of a human heme exporter that is essential for

erythropoiesis. Cell 118:757–766

33. Raffin SB, Woo CH, Roost KT, Price DC, Schmid R (1974)

Intestinal absorption of hemoglobin heme iron cleavage by

mucosal heme oxygenase. J Clin Invest 54:1344–1352

34. Roy CN, Enns CA (2000) Iron homeostasis: new tales from the

crypt. Blood 96:4020–4027

35. Stocker R (1990) Induction of heme oxygenase as a defense

against oxidative stress. Free Radic Res Commun 9:101–112

36. Sun Y, Rotenberg MO, Maines MD (1990) Developmental

expression of heme oxygenase isozymes in rat brain. Two HO-2

mRNAs are detected. J Biol Chem 265:8212–8217

37. Takeda A, Perry G, Abraham NG, Dwyer BE, Kuttyi RK, Lai-

tinen JT, Petersen RB, Smith MA (2000) Overexpression of heme

oxygenase in neuronal cells, the possible interaction with tau.

J Biol Chem 275:5395–5399

38. Tapia N, Arredondo M, Nunez MT (1996) Regulation of Fe

absorption by cultured intestinal epithelia (Caco-2) cell mono-

layers with varied Fe status. Am J Physiol 1271:G443–G447

39. Trakshel GM, Kutty RK, Maines MD (1986) Purification

and characterization of the major constitutive form of testicular

heme oxygenase. The noninducible isoform. J Biol Chem

261:11131–11137

40. Weintraub LR, Weinstein MB, Huser HJ, Rafal S (1968)

Absorption of hemoglobin iron: the role of a heme-splitting

substance in the intestinal mucosa. J Clin Invest 47:531–539

41. Wheby MS, Spyker DA (1981) Hemoglobin iron absorption

kinetics in the iron-deficient dog. Am J Clin Nutr 34:1686–1693

42. Wyllie JC, Kaufman MD (1982) An electron microscopy study of

heme uptake by rat duodenum. Lab Invest 47:471–476

Eur J Nutr (2011) 50:363–371 371

123